Finite dimensional functional spaces for image processing

Abstract:

A polynomial f(x) being shifted becomes a new polynomial f(x+h).
The same is valid for a harmonic and an exponent function.
More generally, a property to be a quasi-polynomial is independent of
shift.
It is obvious and used without stating as a theorem.

In the report we present careful examination of this property that leads
to new point of view for moment invariants and motion estimation.
Examples of pattern recognition are presented.